Chart MPL

Generate PNG/SVG charts (line, bar, hbar, pie, stacked, scatter, area) from CSV data using matplotlib. Use when the user asks to visualize tabular data, prod...

Security Scan

The skill's stated purpose (generate charts from CSV with matplotlib) matches the files and instructions, but the main script was truncated in the package listing so the full runtime behavior couldn't be audited; proceed only after reviewing the complete script or running it in an isolated environment.

Purpose & Capability

Name/description, required binary (python3), SKILL.md usage, and included scripts all align: the skill provides a Python script that reads CSV and renders charts with matplotlib — this is coherent and expected.

Instruction Scope

SKILL.md instructs creating a venv in ~/.openclaw/workspace/.venv_chart, installing matplotlib, and running the bundled script against a provided CSV. Those steps are within scope. It does write outputs to ~/.openclaw/workspace/exports/images (or to a user-specified --out). The audit is limited because the provided scripts/chart_mpl.py was truncated in the package listing, so the latter part of runtime logic (after beginning of scatter branch and any remaining code) could not be inspected.

Install Mechanism

No install spec is present (instruction-only); the SKILL.md recommends a venv and pip install matplotlib. There is no external download of code — the script is included in the package — so install risk is just the normal PyPI install of matplotlib and its dependencies.

Credentials

The skill requests no environment variables or credentials. It reads CSV files supplied by the user and writes image files to the workspace or a user-provided --out path, which is proportionate to the stated purpose.

Persistence & Privilege

always:false and user-invocable:true. The skill creates its own venv and output directories under ~/.openclaw/workspace, which is a reasonable level of persistence and scoped to its own files; it does not request elevated platform privileges.

What to consider before installing

This skill appears to do what it claims (create charts from CSV) and relies on a bundled Python script plus matplotlib. However, the provided script content was truncated in the package listing, so you should not run it unreviewed. Before installing or running: (1) review the full scripts/chart_mpl.py file for any subprocess, networking, or filesystem operations beyond reading the CSV and writing the image (search for os.system, subprocess, socket, requests, urllib, open to unexpected paths, or code that posts files); (2) run pip installs inside an isolated virtual environment or sandbox; (3) if you will supply CSVs exported from other skills, ensure those CSVs don't contain sensitive data you don't want processed or written to disk; (4) consider running the script on a non-production machine first. If you can provide the complete script (no truncation), I can re-evaluate and raise confidence to high if nothing suspicious appears.

Generate chart images from CSV files with scripts/chart_mpl.py.

Setup

Create a virtual environment and install matplotlib (one-time):

python3 -m venv ~/.openclaw/workspace/.venv_chart
~/.openclaw/workspace/.venv_chart/bin/pip install matplotlib

Quick start

Run from workspace root:

~/.openclaw/workspace/.venv_chart/bin/python skills/chart-mpl/scripts/chart_mpl.py \
  --csv /path/to/data.csv \
  --xcol Mois \
  --ycol Valeur \
  --kind line \
  --title "Monthly trend" \
  --xlabel "Month" \
  --ylabel "Value"

Output defaults to: ~/.openclaw/workspace/exports/images

Inputs

• --csv path to CSV file
• --xcol column for X axis
• --ycol numeric column(s) for Y axis — comma-separated for multi-series (e.g. Sales,Costs)
• --kind chart type: line, bar, hbar, pie, stacked, scatter, area
• Optional: --delim if auto-detection fails (;, ,, \t, |)
• Optional: --out custom output path (PNG or SVG — format detected from extension)
• Optional: --title, --xlabel, --ylabel
• Optional: --top N show only top N categories by value (bar, hbar, pie, stacked). Multi-series ranks by row total.
• Optional: --sort sort data before plotting: x-asc, x-desc, y-desc, none (default: none)
• Optional: --numfmt number format on Y axis: fr (1,5M) or en (1.5M) — default: fr

Chart types

Kind	Description	Multi-series
line	Line chart with markers	yes
bar	Vertical bar chart (grouped when multi-series)	yes
hbar	Horizontal bar chart (grouped when multi-series)	yes
pie	Pie chart (first Y column only)	no
stacked	Stacked vertical bar (requires 2+ Y)	yes
scatter	Scatter plot	yes
area	Filled area chart	yes

Notes

• Script auto-detects delimiters when possible.
• Script tolerates European and US numeric formats (1 234,56, 1,234.56).
• Non-numeric values in ycol are converted to NaN with a warning.
• Handles BOM-encoded CSV files (utf-8-sig).
• Multi-series charts automatically show a legend and cycle through an 8-color palette.
• Y-axis ticks use human-readable formatting (K, M, G suffixes) with French or English decimals.
• --top N keeps the N highest categories. For pie charts, the remainder is aggregated into an "Other" slice. For bar/hbar/stacked, the rest are excluded.
• Output format (PNG or SVG) is detected from the --out file extension. Default is PNG.

End-to-end flow with SQL skill

1. Export query result to CSV using the mssql skill.
2. Run this script on that CSV.
3. Share the generated PNG path.